JP6390858B2 - Apparatus and method for comparing two data including graphic element and text element - Google Patents

Description

  The present invention relates to an apparatus for comparing two data including a graphic element and a text element.

  Corresponding devices are known from the prior art as text verification tools (TVT; text verification tools) or object verification tools (OVT). Here, in the case of OVT, the object comparison is performed by a pixel-by-pixel comparison of the two data rendered for it in advance. This gives very good results if this kind of data is a pure graphic component. This is because each different pixel is directly identified as an error and can be displayed. This type of OVT has proven unsuitable for text comparisons. Because when the text is modified, a line break occurs or the character spacing or word spacing changes, which is the difference between the pixels directly, even though no modification has been made. This is because it is indicated as an error. This does not occur with text verification. Here, text is first extracted from each data, and for this purpose, for example, a publicly known OCR program can be used. Similarly, for text identification, text can be first extracted by extraction from complex data such as pdf data, PostScript data, word data with corresponding components defined as text. Furthermore, TVT can even adapt columns or text flows depending on the data format. This is especially true if the text spans multiple lines, or if the text or paragraph number or line number is set before the text, or if a header or footer is placed above or below the page number, for example Is necessary to enable content and textual comparisons. However, this allows text corrections to be identified and directed quickly and accurately by this type of text comparison tool. However, this type of text comparison tool has the disadvantage that graphics cannot be compared. Therefore, finally, it is only possible to make an overall comparison, that is, whether or not the same image exists.

  Thus, the two methods result in subordinate error blockage, so that the correct error can no longer be identified. Correspondingly, the object of the invention is to avoid this drawback in the method and apparatus according to the superordinate concept.

  As a solution, an apparatus or method with the features of the independent claims is proposed.

  Accordingly, an apparatus for comparing two data including a graphic element and a text element includes a memory for the two data, a means for dividing the two data into a graphic element and a text element, and for text identification. Two means: a raster graphic means for forming a rasterized image from graphic elements; a means for comparing extracted text; a means for comparing rasterized images; Means for preferably outputting the results in common.

  A method for comparing two data including a graphic element and a text element also divides the graphic element and the text element in the two data, and forms a raster image from the graphic element as the case may be. It can be characterized by comparing graphically, possibly identifying each text from a text element, and comparing the text textually to each other.

  Therefore, by dividing in advance into graphic elements and text elements, it is possible for the first time to make an appropriate examination or an appropriate comparison for each element separately. This has the advantage that the comparison of correspondences also leads to correspondingly meaningful results and results to be meaningfully interpreted.

  It will be appreciated that the formation of a raster image is only necessary if the corresponding data does not already have a raster format. Similarly, if the data already exists as pure form text, the corresponding text identification can be omitted. However, this is rather an exception, especially in the case of normal professional work, for example in the production of galley and its discussion. This is because the data used there is often in a graphic container format, such as pdf data or PostScript data.

  In order to separate graphic and text elements, it is basically advantageous to first separate the text elements and preferably put them on separate levels. This can be done, for example, by applying a text identification program, such as OCR, to the data present in raster format. The text identified here is correspondingly placed at a separate level and is preferably removed from the graphic level. It is possible to create a level that contains only graphics, possibly in other ways, for example by migrating entire elements that are not text elements to the corresponding graphic level. If the text exists in a graphic container format, for example, a graphic container format region labeled as a text element can be flagged accordingly and subsequently placed on a separate text level.

  Direct text identification continues. For this purpose, glyphs that exist in the form of graphic containers may be assigned to characters via a table known per se, thereby identifying the text as it, and subsequently by text flow to make the text into nro text (natural human reading order). Convert. As a result, the text is connected to a unity which is meaningful in terms of a column format or a table format. This is already well known from the prior art in TVT. In particular, it will be understood that the reading direction or progress direction of this text can be adapted to each language, for example left to right and top to bottom, top to bottom and then for the first time also right to left. Changes in text that may occur in some cases are understood and displayed in a manner that is directly understandable and understandable for humans.

  Therefore, it is also advantageous that the corresponding text identification means includes means for extracting text from the text element and means for displaying the text flow.

  Here, the text identification can be either text at the text level or text at a separate nhro text level.

  It will be appreciated that in corresponding container formats that already contain separate text, complex text identification can be omitted by extracting these texts from their respective container formats as expected. In practice, however, at least one of the data to be compared does not exist in this type of format, or in some cases exists in this type of format but does not contain separate text, and thus the complexities described above Text identification should be performed first.

  If the graphic element does not already exist in raster form, it is advantageous to render it into a raster image, ie convert it to a raster image. It will be understood here that the corresponding raster image can optionally be displayed at the main level of the respective data, or at the graphic level, or at an additional raster image level.

  The raster graphic means for rendering can include a raster graphic engine and / or a raster graphic processor, among others. Accordingly, corresponding rasterization or corresponding rendering is executed at high speed and with a stable operation with a desired accuracy.

  Preferably, the comparison results are displayed in common. This can be done, for example, by superimposing corresponding levels. In particular, the comparison result can be output to a monitor as output means.

  Similarly, it is also conceivable to store the comparison results in data, preferably in common. This data can then be provided, for example, to a customer, for example, to verify corrections made by galling. It is therefore advantageous if the output means includes a memory and storage means for storing the comparison results in result data in the memory.

  If the storage means stores the result data in the form of at least one of the two data to be compared, the comparison results are particularly advantageously further processed. This makes it possible to provide the person who has prepared at least one of the data to be compared with the program to which he belongs in order to open and optionally evaluate the result data.

  It will be appreciated that the features of the solutions recited in the claims can be combined in some cases so that the advantages can be realized correspondingly cumulatively.

  Further advantages, objects and characteristics of the present invention will be explained on the basis of the following description of examples. Examples are shown in particular in the accompanying drawings.

FIG. 1 is a diagram showing a method configuration of the present invention in a lower region and a conventional method configuration in an upper region. FIG. 2 is a diagram showing two data to be compared including a graphic element and a text element, and a comparison result by conventional object verification. FIG. 3 is a diagram illustrating a method configuration according to the present invention and the results based on the data to be compared in FIG. FIG. 4 is a diagram showing a print image of pdf data including a graphic element and a text element as an example. FIG. 5 is a diagram showing the structure of data shown in the print image of FIG. FIG. 6 is a diagram showing the data shown in FIGS. 4 and 5 in a text format. FIG. 7 is a diagram illustrating a comparative example of data including an Arabic text element. FIG. 8 is a diagram showing a comparative example of data including Chinese text elements. FIG. 9 is a diagram illustrating a comparative example of data including graphic elements.

  First, separate text comparisons and separate graphic comparisons are performed by dividing the data 10 and 20 to be compared having graphic elements and text elements separated into graphic levels 11 and 21 and text levels 12 and 22, respectively. Can do. This leads to reasonably reasonable and understandable results. In some cases, in the first step, text identification such as OCR can be applied to the two data 10, 20 to be compared or to only one of these data 10, 20. Or you can search for objects that contain text, or take other steps to create objects that contain text. In this way, firstly, data 10 and 20 in which text elements are also found in a container format from pure graphic data 10 and 20 are provided.

  Graphic levels 11 and 21 are rendered and compared after separation. The result of the graphic comparison is then displayed on a separate level 32.

  In the text levels 12 and 22, text identification and text flow are first executed. Thus, the respective texts 13 and 23 are used for text comparison, and the result is displayed at level 31.

  In this way, document comparisons, especially data comparisons, can be performed in a fast and stable manner, especially if this is a galley to make customer-specific correction wishes under the correction document record.

  By superimposing the levels 31 and 32, it is possible to easily create an entire image by inverting the image 32 in which the corresponding errors are integrally recorded in the document.

  The latter is not possible in the case of object verification as shown in FIG. This is because there is an error blockage. Similarly, it is readily understood that text validation cannot enable graphic comparisons.

  Here, the data can be transferred to a container format such as pdf, and the text element 43 and the graphic element 44 can be found in this type of format. For example, if the data shown in the printed image of FIG. 4 has a circle as the graphic element 44 and “Hello World” as the text element 43, the text can be found separately from the graphic, and in particular this It can be read from the data structure display. If it is not always clear whether the corresponding information is also included in the textual representation of the respective data (see FIG. 6), the text element is mechanically separated from the graphic element accordingly and shown above. Can be inspected as follows.

  It is obvious that the text element is not necessarily limited to Roman letters. Other character languages such as Arabic or Chinese can also be investigated and compared correspondingly. This is illustrated by way of example in FIGS. 7 and 8, where the reading direction is adapted according to the language or typeface. However, this is usually not a problem by embedding the text accordingly. Again, if appropriate, the corresponding missing or errored part 45 can be marked, so that a simple and stable comparison can be made.

  Errored or missing portions 45 in the graphic can be marked accordingly. This is shown as an example in FIG.

10 data 11 graphic level 12 text level 13 text 20 data 21 graphic level 22 text level 23 text 31 level for text comparison result 32 level for graphic comparison result 43 text area for data 44 graphic area for data 45 missing part or Error part

Claims (11)

In an apparatus for comparing two data including a graphic element and a text element and outputting a difference, a memory for the two data, means for dividing the two data into a graphic element and a text element, text Means for identifying; raster graphic means for forming a rasterized image from graphic elements; means for comparing extracted text; means for comparing rasterized images; To display the difference between the extracted text and the rasterized image on the same screen by superimposing the level indicating the difference between the text and the level indicating the difference between the rasterized images of the two data. And an output means.   The apparatus of claim 1, wherein the raster graphic means comprises a raster graphic engine and / or a raster graphic processor. 3. An apparatus according to claim 1 or 2 , wherein said output means includes a memory and storage means for storing text differences and image differences in result data in the memory. 4. The apparatus according to claim 3 , wherein the storage means stores the text difference data and the image difference data in the form of at least one of two data to be compared. Apparatus according to any one of claims 1 to 4 wherein the output means including a monitor, it is characterized. It said text identifying means, means for extracting the text from the text element, and means for displaying the text flow, that the device according to any one of claims 1 to 5, characterized in. In a method for comparing two data including a graphic element and a text element and outputting a difference, the graphic element and the text element are separated from each other in the two data, and if necessary, one raster image is obtained from each graphic element. Graphically compare raster images with each other, identify each text if necessary from text elements, compare text textually with each other, and show differences in levels and text indicating raster image differences A method of displaying a difference between raster images of two data and a difference between identified texts on the same screen by superimposing levels . The method according to claim 7 , wherein the raster image difference and the text difference are output to one file. 9. A method according to claim 7 or 8 , wherein at least one of the graphic elements is rendered into a raster image. 10. A method according to any one of claims 7 to 9 , characterized in that the text elements are first placed on separate levels in order to separate the graphic elements and the text elements. 11. A method according to any one of claims 7 to 10 , characterized in that for text identification, text in a form that is easy to read is formed by a text flow.